HOW MIGHT SEDIMENT SIZE HETEROGENEITY INHIBIT RIVER INCISION?

Rapidly eroding rivers influence adjacent hillslopes by setting their lower boundary conditions, while hillslopes and canyon walls affect rivers by supplying sediment of varying size and lithology. Much progress has been made in identifying the influence of sediment mass flux on river incision, and specifically in identifying the prevalence of erosion-inhibiting bed armor at high sediment supplies. Sediment supplied to rapidly eroding rivers is often very heterogeneous in size, and we argue that significant bed armoring by the largest grains may occur even if sediment mass flux does not exceed a river's carrying capacity. This leads to cases where rivers that quickly cut steep-walled canyons incite a negative feedback on incision by receiving an increased supply of large, immobile sediment from the canyon walls that shields significant portions of the bed from erosion. We use a simple analytical model to demonstrate that this effect is controlled by the sediment size distribution supplied from the hillslopes, the input rate of new large grains, and the rate at which large grains degrade. In model cases where the channel is decoupled from the hillslopes (i.e., increased incision does not increase the supply of large grains), the pace of river incision remains consistent through time. In cases where rapid river incision leads to an increased influx of large grains from the hillslopes (reflecting the hypothesis that steeper slopes will be more prone to releasing large grains), erosion slows as the fraction of the bed shielded by immobile grains increases. This result suggests an inherent negative feedback on incision in channels with access to a supply of large grains. Preliminary results from field surveys in the canyons of the Wind River and Beartooth ranges suggest that the residence time of large boulders and blocks in a single channel reach is long enough to cause significant long-term bed armoring. Our results show that changes in sediment size distribution may exert significant control over rates of river and landscape evolution. The critical governing timescales of this effect are the input rate of large, immobile grains from the hillslopes and the degradation rate of large grains in the channel, both of which warrant further investigation.